BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250626T234217Z
LOCATION:B308
DTSTART;TZID=America/New_York:20241120T133000
DTEND;TZID=America/New_York:20241120T150000
UID:submissions.supercomputing.org_SC24_sess397@linklings.com
SUMMARY:Machine Learning Applications
DESCRIPTION:Scaling New Heights: Transformative Cross-GPU Sampling for Tra
 ining Billion-Edge Graphs\n\nTraining GNNs on billion-edge graphs faces si
 gnificant memory and data transfer bottlenecks, especially with GPU-based 
 sampling. Traditional methods struggle with CPU-GPU data transfer bottlene
 cks or high data shuffling and synchronization overheads in multi-GPU setu
 ps. \nTo overcome these challenges...\n\n\nYaqi Xia (Wuhan University Scho
 ol of Computer Science); Donglin Yang (NVIDIA Corporation); Xiaobo Zhou (I
 OTSC, University of Macau); and Dazhao Cheng (Wuhan University School of C
 omputer Science)\n---------------------\nAccelerating Distributed DLRM Tra
 ining with Optimized TT Decomposition and Micro-Batching\n\nDeep Learning 
 Recommendation Models (DLRMs) face challenges due to the high memory needs
  of embedding tables and significant communication overhead in distributed
  settings. Traditional methods, like Tensor-Train (TT) decomposition, comp
 ress these tables effectively but add computational load. Furthe...\n\n\nW
 eihu Wang and Yaqi Xia (Wuhan University School of Computer Science); Dong
 lin Yang (NVIDIA Corporation); Xiaobo Zhou (IOTSC & Department of CIS, Uni
 versity of Macau); and Dazhao Cheng (Wuhan University School of Computer S
 cience)\n---------------------\nA Scalable Algorithm for Active Learning\n
 \nFIRAL is a recently proposed deterministic active learning algorithm for
  multiclass classification using logistic regression. It was shown to outp
 erform the state-of-the-art in terms of accuracy and robustness and comes 
 with theoretical performance guarantees. However, its scalability suffers 
 when d...\n\n\nYouguang Chen, Zheyu Wen, and George Biros (The University 
 of Texas at Austin)\n\nTag: Accelerators, Applications and Application Fra
 meworks, Distributed Computing, Graph Algorithms, Heterogeneous Computing,
  Tensors\n\nRegistration Category: Tech Program Reg Pass\n\nSession Chair:
  Mohamed Wahib Attia (Graduate School of Info. Sccience and Tech., Hokkaid
 o University, Sapporo, Japan)
END:VEVENT
END:VCALENDAR